AIR1-CT92-0513 ASHLI - Fundamental Studies on the Selective Hydrogenation of Fats and Oils

Contacts
Summary Information



To find similar Items, click on a keyword below:
AIR Cluster IV - Oils and Fats : Bulk Chemicals : Chemical Conversion : Process Engineering : Vegetable Oil/Fat

	Proposal No:	AIR1-CT92-0513
	Date Prepared:	September 1999
	Source:	Final report 1998

---

Final report 1998

Introduction
Increase of the degree of saturation of fatty oils via hydrogenation is the most important process of the fatty oil industry, particularly in the production of edible fat products. Industrially, the process is performed by passing hydrogen gas through a stirred slurry of the oil and supported nickel catalyst. Though already a classical process, widely applied since early this century, it is not yet possible to a priori predict the molecular composition of the hydrogenated oil as a function of feed stock composition, catalyst type and concentration, reaction pressure, temperature and time. This is partly due to the complexity of the process, with a large number of hydrogenation reactions occurring in parallel, and partly due to the simultaneous occurrence of cis/trans isomerisation and double bond conjugation reactions.

Some of the hydrogenation reactions occurring on nickel catalysts are less desirable. Especially in some food products, trans isomers are considered undesirable. An important factor in the performance of the hydrogenation is therefore the selectivity regarding the suppression of the formation of trans isomers.

Objectives
Such considerations resulted in the following objectives of the ASHLI-project:

• ability to hydrogenate fats and oils in such a way that product composition can be varied in a controlled and predictable way, based upon a reaction engineering model for supported nickel catalyst
• investigation into the possibility to substantially improve the selectivity of the hydrogenation of edible oils with regards to formation of trans isomers
• investigation into the possibility to hydrogenate rapeseed oils with substantially reduced formation of trans isomers by applying improved hydrogenation techniques

Activities
The research performed to obtain a better understanding of the nickel-catalysed hydrogenation process, and to adequately model the nickel-catalysed hydrogenation process, included the following subjects:

• measurement of the intrinsic chemical kinetics of nickel-catalysed hydrogenation in the absence of pore diffusion limitation
• measurement the effective pore diffusion coefficients in conventional nickel catalysts
• measurement the adsorption of hydrogen and fatty acid derivatives on the surfaceof the nickel catalyst
• preparation of catalysts with highly regular and uniform pores to investigate the influence of mass transfer on the kinetics
• investigation of opportunities for improved nickel catalysts by applying dopes, surfactant and inhibitors

New reactor and catalyst types were investigated to improve the quality of the hydrogenation, especially with regard to the reduction of the cis/trans isomerisation rate. The following research items were included here:

• development of a standardised method for the evaluation of the performance of catalysts in hydrogenation tests
• preparation and testing of new, precious metal catalyst systems
• investigation into the structural requirements for a good hydrogenation catalyst
• investigation of promising zeolite catalysts for their selective hydrogenation ability
• testing of catalysts for production of hydrogenated fats especially from rapeseed oils
• investigation of reactor systems for optimum application of new catalysts in practice
• investigation of catalyst poisoning and regeneration

Additional work, in order to facilitate the various fields of research mentioned above, included:

• preparation of many specific and purified fatty components
• development of new, highly sophisticated analytical techniques
• design and construction of various specific experimental set-ups at the participating institutes

Results
A standard experimental measuring method was developed for the determination of activity, selectivity and cis/trans isomerisation rate of hydrogenation systems. The standardisation ensured that the true intrinsic kinetics of catalysts are obtained, allowed the results of participating institutes to be compared, and allowed conventional hydrogenation systems and catalyst to be compared with alternative ones.

With a comprehensive experimental programme, the intrinsic kinetics were elucidated for the nickel-catalysed hydrogenation of monoenes, dienes and trienes. The mechanisms of the surface reactions were evaluated from the experimental data using systematically derived sets of rate equations (Langmuir-Hinshelwood-Hougen-Watson approach) and statistical methods (Bartlett test). A new approach was developed, by which the many different types of reactions are modelling as double bond specific, and not fatty acid specific. This new concept in hydrogenation modelling proved very successful in describing the experimental data. The resulting kinetic models are unprecedented in accuracy, in the large number of positional and geometrical isomers described, and in the applicability to a wide range of process conditions.

A complete view of the intra-particle diffusion in porous nickel catalysts was obtained from combining the results of tracer pulse experiments on packed columns and of diffusion controlled hydrogenation measurements with mathematical modelling of the diffusion processes. Effective intra-particle diffusion coefficients of hydrogen and fatty components were obtained. This research project is among the first to successfully apply the tracer pulse technique to very small catalyst particles (10 micron).

The adsorption phenomena of fatty acid derivatives were studied with tracer pulse and batch adsorption experiments. On supported porous nickel catalysts, strong adsorption of the ester functionality of the components onto the carrier material overruled any preferential adsorption of double bonds. With non-supported nickel catalysts, the following ordering of preferential adsorption, although already generally assumed in the literature, was now experimentally established: dienes > monoenes > saturated.

Model catalysts with straight unbranched pores of uniform length and diameter were prepared by anodic oxidation of alumina. With these highly uniform catalysts, a fundamental investigation into the influence of mass transfer on the hydrogenation kinetics was conducted. From the results of hydrogenation experiments with these model catalysts it was shown that catalysts with pore lengths of 2-3 microns are preferable for obtaining high linoleate and cis/trans selectivities.

Following indications from the scientific literature, the possibility was tested of decreasing the cis/trans isomerisation rate of nickel catalyst by addition of certain amine components. A series of hydrogenation experiments was performed with the addition of ammonia, dodecylamine, ethylene diamine and vitamin BI2 to the slurry of nickel catalyst in fatty oil. Compared to the performance without additions, the presence of the amines resulted in lower linoleate selectivities or lower reaction rates. With all amines, the cis/trans ratio was unchanged, which led to the conclusion that application of these additives is not recommendable.

In a comprehensive programme, 150 new types of base metal and precious metal catalysts for partial hydrogenation were prepared and tested. Among the wide range of new catalysts were metal catalysts on many different supports, unsupported metal and alloyed metal catalysts, doped and promoted precious metal catalysts, etc. The development of a specific type of catalyst that showed up to 15 times higher activity and substantially (3.4 times) lower amounts of trans isomers compared to the industrial nickel catalysts in the hydrogenation of sunflower oil was very successful.

The specific type of catalyst that was selected as the most promising, out of the programme of the preparation and testing of new catalysts, was tested in the production of hydrogenated fat from rapeseed oils. From the results of test runs performed at temperatures of 333-363 degrees K and a hydrogen pressure of 0.5 MPa, the following conclusions were reached:

• the observed selectivities and formation of trans isomers with rapeseed oils were similar to those obtained with sunflower oil
• the resulting amounts of trans isomers (5% at 333 Y, 100% at 363 K) in the hydrogenation of rapeseed oils were significantly lower than those obtained with the nickel catalyst (20% at 363 K, 50% at 453 K)
• the linoleate selectivities with the selected new catalyst in rapeseed oil were similar to that obtained with nickel catalyst at the same temperature

To investigate the possibilities of using zeolite catalysts in the selective hydrogenation of vegetable oils, five different types of zeolites were impregnated with 0.01-1.0% palladium as the catalytic material. These were used in hydrogenation experiments of sunflower oil methyl ester under standard ASHLI testing conditions. Unfortunately, in all experiments, a large degree of cis/trans isomerisation was observed, resulting in large amounts of the undesired trans isomers, combined with a low linoleate selectivity.

From a literature study on the types of reactors applied in oil hydrogenation it was concluded that monolith reactors are very promising for application with the new catalysts. Consequently, the performance of monolith catalyst reactors in hydrogenation was extensively tested. A number of advantages of monoliths over conventional slurry reactors were encountered:

• lower pressure drop
• absence of the need for catalyst filtration
• suitability for continuous, steady state operation
• higher productivity per volume of reactor content

A possible problem with the application of monolithic catalyst reactors is the enhanced vulnerability to catalyst poisoning, because the catalyst is fixed in the reactor, compared to the classical slurry reactors. Therefore, various catalytic metals were studied in the monolith reactor for their sensitivity to the deliberate addition of different amounts of sulphur and the long-term deactivation due to coke formation, as well as the possibility to regenerate the catalyst from these poisoning effects. Very promising results were obtained from monoliths impregnated with a specific type of catalyst: a simple procedure was developed, resulted in complete regeneration of this catalyst after sulphur poisoning as well as long-term deactivation.

In order to facilitate various fields of research specific and purified fatty components were prepared. Among these were myristic acid propyl ester, sunflower oil methyl ester, high oleic sunflower oil, and the methyl esters of petroselinic, elaidic, erucic, linoleic, eleostearic, trans-trans-finoleic, and alpha- and gamma-linolenic acid.

The hydrogenation product mixtures sometimes contained up to 30 positional isomers, both cis and trans. For the identification of the components, gas chromatography (GC) was combined with mass spectrometry. For the quantitative analysis a new technique was developed, by combining GC with high performance liquid chromatography. A new method for GC analysis of triacylglycerides was developed, based on fast and complete methylation pre-treatment of the sample before injection on the GC.

Conclusions
A detailed and fundamental scientific investigation into the reaction kinetics of the nickel-catalysed partial hydrogenation was completed. This comprehensive programme included investigations of the intrinsic chemical kinetics, of the effective pore diffusion coefficients, of the adsorption onto the catalyst surface, of the influence of mass transfer on the kinetics using model catalysts with highly uniform pores, and of the opportunities for improving the nickel catalyst by applying additives.

The complex oil hardening process, with an intricate system of simultaneous hydrogenation, cis/trans isomerisation, and double bond conjugation reactions at the surface of the nickel catalyst, previously not well understood, can now be described and modelled accurately. Given the initial oil composition, and the process conditions, the distribution of dozens of different product components, including cis/trans isomers and geometrical isomers, can be predicted.

To investigate the possibility of improving the selectivity as far as formation of trans isomers is concerned, 150 new types metal catalysts for partial hydrogenation were prepared and tested. From this extensive research programme a specific type of catalysts was developed, that showed substantially (3.4 times) lower amounts of trans isomers in the hydrogenation of sunflower oil, when compared to the industrial nickel catalysts, and at the same time showed an up to 15 times higher activity and maintained a similar linoleate selectivity.

This specific new catalyst was tested in the production of hydrogenated fat from rapeseed oils. The observed amounts of trans isomers in the hydrogenation of rapeseed oils with the new catalyst were significantly lower (>4 times) than those obtained with the traditional nickel catalyst while also achieving at least similar linoleate selectivities and much higher activities.

The performance of monolith catalyst reactors in the selective hydrogenation of oils was extensively tested. The catalysts are present in a fixed form inside this type of reactors, resulting in negligible catalyst losses and the absence of the need for filtration of the product. This important advantages of monoliths over conventional slurry reactors, especially in the application of precious metal catalysts, is augmented by the possible problem of an enhanced vulnerability to catalyst poisoning. However, very promising results were obtained from monoliths impregnated with a specific type of catalyst, in that a simple procedure was developed, resulted in complete regeneration of this catalyst after sulphur poisoning as well as long-term deactivation due to coke formation.